Realtime AI News
Manus AI to resume standalone operations after Meta acquisition collapses
Manus AI told users it will soon return to operating as an independent company, ending Meta's attempt to acquire the Chinese agentic AI startup. Chinese regulators blocked foreign investment in Manus in April, and the company will delete some user data on August 24 to satisfy legal requirements.
Manus AI, the Chinese startup behind one of the most prominent “general agent” products, told users it will soon return to operating as an independent company, confirming that Meta's attempt to acquire it has fallen apart. The Register reported the notice, which opens with the line “Manus will soon return to operating as an independent company.”
Manus burst onto the AI scene in March 2025 with a general agent that let users assign tasks to an AI. To illustrate its powers, Manus claimed it could evaluate job applications stored in a ZIP file by opening and reading them, assessing them against user-defined criteria, and writing a document ranking the best applicants — all inside a virtual PC that its agents drove without human help.
The service quickly gathered enthusiastic users and claimed to have hit a $100 million run rate. To fund the GPU purchases an AI company needs, Manus moved its official headquarters from China to investor-friendly Singapore.
In December 2025, Meta founder and CEO Mark Zuckerberg decided to acquire Manus, planning to use its technology across Meta's AI products. “Manus's exceptional talent will join Meta's team to deliver general-purpose agents across our consumer and business products, including in Meta AI,” Meta said in its acquisition announcement.
Beijing had other ideas. Chinese regulators launched an investigation into the acquisition in early 2026, and by April they had decided to forbid foreign investment in Manus. Meta chose not to fight China over the matter and began disentangling itself from the startup.
That effort now appears near complete, but the separation is not painless: Manus says legal matters require it to delete some user data generated since December 29, 2025. The deletion happens on August 24; customers who download their data before then can restore it to Manus from August 25.
Meta has not said how or whether it will replace Manus's technology, or whether its brief time working with the Chinese company yielded what it needed for the superintelligence ambitions Zuckerberg described this week.
The episode underscores China's determination to keep control of local innovators rather than let them leave in search of foreign investment — even as Chinese companies use open-weight models to challenge American model makers.
Why it matters
The collapse of the deal shows China's determination to keep local AI innovators at home, leaving Meta without a marquee agentic AI acquisition and adding another chapter to US-China AI decoupling.
Nearby Updates
All08/12, 13:50
DeepSeek and Chinese tech peers begin building their own AI computing infrastructure
DeepSeek and other Chinese tech firms have started developing their own AI computing infrastructure, Yicai Global reports. The shift toward self-built compute gives leading AI labs more control over cost and capacity amid constrained access to advanced accelerators.
08/12, 13:55
Alibaba Cloud makes its M890 AI supernode available in China
Alibaba Cloud has made its M890 AI supernode available in China, according to a report from TechNode. The large-scale, purpose-built AI compute cluster gives domestic customers model-scale capacity for training and inference without assembling the hardware themselves.
08/12, 11:51
Dongyin Technology launches cross-border e-commerce TMS driven by AI Agent
Dongyin Technology (洞隐科技) has officially launched a cross-border e-commerce transportation management system (TMS) driven by AI agents, aimed at connecting the full outbound logistics chain. Reported by Sohu, the release signals agentic AI moving deeper into vertical logistics workflows.
08/12, 14:00
OpenAI enterprise report: Codex now drives 64% of enterprise output tokens as agentic AI spreads
OpenAI published two studies showing enterprise AI is shifting from assistants to agents: Codex generated 64% of combined Codex and ChatGPT output tokens among enterprise customers as of June. Frontier firms now produce 8.3x more output tokens per active user than typical firms, up from 2.6x in January, with agentic adoption spreading fast across legal, sales, recruiting, and marketing.